after contrail-api restarted on tor-scale setup, agent didnt get any xmpp config from control node
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R2.20 |
Fix Committed
|
High
|
Tapan Karwa | |||
Trunk |
Fix Committed
|
High
|
Tapan Karwa |
Bug Description
R2.20 30 Ubuntu 14.04 Juno multi-node setup
env.roledefs = {
'all': [host2, host3, host4, host5, host6, host7, host8, host9],
'cfgm': [host2, host3, host4],
'openstack': [host2, host3, host4],
'webui': [host3],
'control': [host2, host3, host4],
'compute': [host5, host6, host7, host8, host9],
'collector': [host2, host3, host4],
'database': [host2, host3, host4],
'toragent': [host5, host6, host7, host9 ],
'tsn': [host5, host6, host7,host9 ],
'build': [host_build],
}
env.hostnames = {
'all': ['nodei34', 'nodei35', 'nodei36', 'nodei37', 'nodei38', 'nodei28', 'nodei27', 'nodei30']
}
There are 128 tors each with 1k vmis and 512 lifs
For some reason today, contrail-api restarted on all the 3 config nodes around 11:45 AM .
It was later(after ~5-10 mins) seen that control-node had deleted the xmpp config from all agents/tor-agents, but never added it back
There were no xmpp flaps during this period
Logs from all 3 nodes will be in http://
gcore of control-node and agent of nodei27(a compute-node ) will be in the same location
Changed in juniperopenstack: | |
assignee: | nobody → Tapan Karwa (tkarwa) |
introspect detail